INTERNATIONAL TELECOMMUNICATION UNION TELECOMMUNICATION STANDARDIZATION SECTOR STUDY GROUP 3 TD 31 (PLEN/3) - E Original: English STUDY PERIOD 2009-2012 Question(s): ALL/3 Geneva, 19-23 January 2009 TEMPORARY DOCUMENT Source: TSB Title: Tutorial on Basic Economics and Telecommunications: Regulating Telecommunications Services This note, to accompany an ITU presentation in January 2009, reviews the way in which telecommunications and related services are regulated, and the new challenges created by technological and other developments such broadband and the increased use of internet-based technologies. The focus is both on fixed and on mobile service, with a greater emphasis on the former because they present greater problems of regulation. 1. Reasons to regulate telecommunications services Until the 1980s, there was often unthinking acceptance that telecommunications services required regulation because they were based on natural monopoly infrastructures. The North American model for dealing with this supposed attribute was via regulation of investor-owned enterprises, usually on a cost-plus (rate of return) basis (Brock, 2002). The ‘European’ model, widely followed elsewhere, rested upon public ownership, with services delivered through a government department or a company wholly owned by the government in question. The absence of competition in either model allowed a range of regulatory objectives to be delivered relating to the availability of services and the terms and conditions of their supply. There was also no difficulty in principle in ensuring that the industry covered its costs: the monopoly firm could simple raise prices to do so. The US model of rate of return or cost-plus regulation had precisely that objective and effect. Almost any configuration of tariffs could work. However, the introduction of competition in many parts of the industry, accompanied by privatisation, forced a more rigorous analysis of potential market failures and evinced a radical regulatory response, and this has been accompanied by a clearer articulation of the objectives and instruments of non-economic or social regulation, directed primarily at ensuring universal service at affordable prices. Contact: Martin Cave Tel: +44 7958 483709 Fax: +44 24 7652 4965 Email martin.cave@wbs.ac.uk Contact: Richard Hill Tel: +41 22 730 5887 Fax: +41 22 730 5853 Email richard.hill@itu.int Attention: This is not a publication made available to the public, but an internal ITU-T Document intended only for use by the Member States of ITU, by ITU-T Sector Members and Associates, and their respective staff and collaborators in their ITU related work. It shall not be made available to, and used by, any other persons or entities without the prior written consent of ITU-T. -2TD 31 (PLEN/3) In summary, telecommunications regulation can be seen as addressing two types of problems: 1. 2. market failure, associated with high levels of monopolisation deriving from:1 - economies of scale (costs falling as output increases) - economies of density (associated particularly with access networks) - economies of scope (when two services, such as telecommunications and broadcasting, are provided more cheaply over a single network) - demand-side network externalities (where customers derive greater benefits from belonging to a network with more, rather than fewer members) non-economic objectives, notably - universal service, ensuring that service is available everywhere at a uniform price, and redistributive objectives, designed to protect, for example, low income households or people with disabilities, and political inclusion. The alternative to these outcomes in the digital age is often captured by the phrase ‘digital divide’. Turning first to market failures due to monopolisation, it has proved exceptionally difficult to replicate the fixed access network. The exceptions are the use of cable TV networks or wireless via networks whose outputs in mobile form do not (according to most regulators) compete directly with fixed calls. Experience suggests that retailing, or reselling the incumbent’s products, is wholly contestable,2 while the copper access loop is practically impossible to replicate – although cable systems and wireless access networks offer competitive possibilities in some contexts. Other activities such as backhaul from local to trunk exchanges, and the high capacity transport among such trunk exchanges fall into an intermediate position.3 It follows from this that, as competition develops, entrants may progressively install some capacity, but rely on the incumbent to supply the rest. Thus they may start from retailing, that is, reselling the incumbent’s services, and progressing via the installation of a core network connecting a small number of trunk switches, and later extend into backhaul and the replication of the incumbent’s local exchange assets. Section 3 below discusses an approach to regulation of broadband based that facilitates this gradual process of infrastructure competition. Economies of scope play an increasing role in telecommunications as a result of technological developments, especially digitisation. Whereas broadcasting, voice telecommunications and computer-based data networks used to exist in separate silos, they have now ‘converged’ and the relevant information or bits for each is carried indistinguishably on a variety of networks. Thus cable networks offer the ‘triple play’ of voice, broadcast services and broadband. Telecommunications networks based on DSL (digital subscriber lines) technologies can do the same, provided they have the capacity to convey video services, as do the ‘next generation’ networks described below. Increasingly wireless networks, static, mobile or nomadic, can offer similar combinations. Satellite operators can combine with wire-based return paths to offer a range of services. As a result, markets have widened, creating the scope for new competitive opportunities and for new practices such as the bundling of services by dominant operators in ways which limit or distort competition. 1 For evidence of the cost characteristics of fixed telecommunications networks, see Fuss and Waverman (2002) and Sharkey (2002). 2 The concept of ‘contestability’ derives from the work of Baumol and others in the 1980s, and refers to the idea that low barriers to entry pressurise the incumbent to act in a way that mimics competition. 3 Some operators are starting to introduce 'next generation networks’ (NGNs). See Section 3 below. -3TD 31 (PLEN/3) The final potential source of market failure noted above arises from the demand-side network effects associated with communications networks. The number of potential interchanges between network members grows with the square of their number.4 Clearly, without interconnection of networks, there would be a tendency either for customers to ‘multi-home’, that is subscribe to many networks (which would be expensive) or for one network (the largest) to drive all others out. However this danger can be averted, and the benefits of ‘any-to-any connectivity’ can be gained, by mandating interconnection.5 The non-economic objectives of regulation noted above require a different approach.6 In essence, policy makers have imposed on regulators objectives which go beyond avoiding market failure and replicating, through regulation, the outcome of a competitive market process. When telecommunication was a monopoly, non-economic objectives could be pursued by cross-subsidy, for example by charging the same prices in low cost and high cost areas. But when competition is present, no-operator will want to serve high-cost customers, if they can only charge an average price. All operators will seek to ‘cherry-pick’ low cost areas. The resulting stresses created and ways to overcome them are discussed below. 2. Traditional Regulatory Approaches and their Reform This section describes is a stylised way the way in which regulators have over the past 20 years addressed the regulation of fixed voice services.7 In a liberalised environment this involves three particular forms of intervention, retail price control, the regulation of access and interconnection, and universal service. Three stages of market structure are distinguished, characterised as ‘monopoly’, ‘transition’ and ‘normalisation’. The last is a stage where most markets, apart from bottlenecks zones such as call termination, have been opened up to competition. Transition is the rather elastic period between monopoly and normalisation. As the account which follows makes clear, the third stage has proved elusive to date, but it remains a useful target for the design of regulation.8 The key structural break occurs when entry into fixed networks and services is liberalised. This has taken place at various dates over the past 20 years in most countries.9 It should be pointed out that apparent liberalisation of entry can be deceptive, if the government or regulatory authority is seeking to maintain barriers to entry by imposing unnecessarily onerous licensing obligations. 4 Though the value of those interchanges does not, as we are less interested in talking to perfect strangers than to friends and relations. 5 This consequence applies not only to voice services, but also to, for example, instant messaging and the sharing of content on networks, via companies such as My Space. The ‘network effects’ problem has surfaced in another form in mobile networks. Mobile operators typically offer their customers lower price for on-net calls (to a mobile subscriber on the same network) than for off-net calls (to another mobile network). This can enhance the attractiveness of belonging to the largest networks, or to one chosen by one’s peer group. There has been debate about whether such prices are discriminatory or anti-competitive, but no adverse regulatory decision about such so-called ‘tariff-mediated network externalities’ has so far been taken. 6 Universal service may, of course, have economic as well as social objectives, as the spread of communication services can spill over into the economy more broadly. 7 For a detailed account of developments in the sector and its regulation, see ITU (2004). 8 Thus the regulatory regime operating in the European Union since 2003, permits the application of sector-specific rules only where there is a dominant firm operating, in a market characterised by persistent barriers to entry or development of competition. Absent these conditions, regulation is unlawful. (See Buigues and Rey, 2004). 9 Entry into mobile services- or wireless services more generally – has been limited by the availability of spectrum. But most larger markets have had enough licences created to achieve something approximating ‘workable competition’, although the barriers to entry do encourage tacitly collusive prices. -4TD 31 (PLEN/3) As far as behavioural regulation is concerned, three behavioural instruments are required in the early stages of liberalisation: 1. control of retail prices is necessary where the dominant firm exercises market power at the retail level, since in the absence of retail price control, customers will be significantly disadvantaged. However, as competition develops at the retail level, possibly from firms relying largely on infrastructure belonging to the incumbent, the necessity for retail price controls in effectively competitive markets may disappear, although access price control may still be necessary. 2 in order to maintain any-to-any connectivity in the presence of competitive networks, operators require interconnection to one another’s networks in order to complete their customers’ calls. This requires the operation of a system of inter-operator wholesale or network access prices. Especially in the early stages of competition, entrants will require significant access to the dominant incumbent’s network, and this relationship will almost inevitably necessitate regulatory intervention. However, as infrastructure is duplicated the need for direct price regulation of certain network assets diminishes. 3 governments have typically imposed a universal service obligation on the historic telecommunications operator, based upon two requirements: an obligation to provide service to all parts of the country, and to provide at a uniform price, despite the presence of significant cost differences. Entrants coming into the market without such an obligation have a strong incentive to focus upon low-cost, ‘profitable’ customers, putting the USO operator at a disadvantage. Pressure may therefore build up to equalise the situation, perhaps by calculating the net cost of the USO borne by the dominant operator in serving loss making customers and then sharing the cost among all operators. There has been concern, first that such an arrangement would be used as a pretext for delaying competition, secondly that high USO contributions imposed on entrants would choke off competitors. In practice, most regulators, with the notable exception being Australia, have maintained the USO as an obligation on the incumbent, without introducing cost sharing. Table1: Stages of regulation Monopoly Transition Normalisation Retail price control Controls on all services Gradual relaxation of controls No controls Access pricing Not relevant, or arbitrary pricing of small range of services Services gradually decontrolled Unbundled Use of price caps Call termination only controlled Universal service obligations Borne by incumbent Costed and shared (or ignored if not material) As in ‘transition’, with possibility of competitive provision Regulatory policy in relation to the control of retail and wholesale prices and the allocation of universal service obligations is likely to change throughout the three phases of liberalisation: monopoly, transition and normalization. Table 1 sets out the way in which each of the three regulatory instruments might develop. -5TD 31 (PLEN/3) Interconnection Pricing and other Conditions10 Access to networks is of fundamental importance for the development of effective competition. The telecommunications sector from entrants need to interconnect with established network operators. The price which entrants have to pay for access to the incumbent network is crucial to their commercial success, yet the incumbent has two motives for charging high access prices. The first is a simple desire to maximise monopoly profits; the second, which arises if a network owner is also competing in the retail market, is the desire to raise its rivals' costs and maintain a dominant position in that market. These considerations have meant that regulators are likely to have to intervene in access pricing either by imposing detailed prices for the use of the elements of the network, or by limiting the overall revenues which the network owner can collect for providing its services to those which are necessary for the recovery of its costs. Either approach involves a detailed analysis by the regulator of the costs incurred by the network. The first-best solution would involve setting access prices on the basis of marginal cost. This may need to be supplemented by a mark up if such pricing would prevent the incumbent from breaking even. The standard method of interconnection and access pricing is thus to set price based on costs (more particularly on long run incremental costs, with a mark up to cover common costs.11 In Europe the 1997 Interconnection Directive required that charges for interconnection follow the principles of transparency and cost orientation, and this has broadly been maintained under the subsequent regulatory regime (Buiges and Rey 2004). The first principle implies the publication of a reference interconnection offer (RIO). As a corollary, operators with significant market power are usually required to keep separate accounts for their wholesale or network activity and for other activities, including retailing. Universal Service The notion that a country’s telecommunications operator should have a universal service obligation has a long history, becoming well established in the period of monopoly provision by either a public or a private operator. In telecommunications, universal service is widely recognized as having at least four dimensions: the achievement of universal geographical coverage, requiring that a network is rolled out over the whole of the territory in question, so that service is available for the whole population. geographic averaging of tariffs, requiring that customers in the same category (for example domestic or business customers), are charged the same tariffs, irrespective of their individual costs of supply. a basic telecommunication service should be made available at an 'affordable' price, so that no household, or more realistically only a few households, are denied access to the service. This is increasingly replaced by the more nuanced proposition that certain households, characterised by attributes such as low income or by their pattern of use of the telecommunications network should have special 'social' tariffs available to them. more recently, discussion has turned to whether broadband should be included in the products provided as a universal service. At current levels of broadband penetration, this is probably premature in most countries. 10 11 Many other forms of asset sharing are considered in ITU (2008). See OECD (2004) for a review. -6TD 31 (PLEN/3) The rationale for having a universal service obligation is based on a mixture of political, social and economic considerations. It is desirable on political grounds that citizens in a democracy have access to communication facilities which they require to exercise their political rights, and it is desirable on social grounds that all individuals should have access to communication facilities, to avoid a ‘digital divide’ emerging between 'information-rich' and 'information-poor' groups. Economic arguments derive from 'network externalities' and the link between network expansion and economic growth. The notion that a universal service obligation for telecommunications should be sustained has survived the introduction of competition in the sector. It had previously been thought that competition and the USO were incompatible, because competitors would focus upon high value customers, defined either by their calling patterns or their (low cost) place of residence, leaving low value customers to the operator with the universal service obligation. As competitors aggressively encroached upon profitable markets, the commercial prospects of the USO operator would gradually unravel, jeopardising the universal service obligation. However, further examination of the issue quickly revealed that it was possible to avoid this outcome if the cost of the universal service obligation were shared on some equitable basis between the operator discharging it and competitors. This triggered a discussion of appropriate means of estimating the cost and appropriate means of sharing it, and a number of the governments and regulatory agencies have implemented procedures estimating the cost, and in some cases have instituted arrangements for sharing the costs. These developments have encouraged a new view of the universal service obligation and its funding as a sui generis tax and subsidy regime. Effectively, a limited amount of revenue is raised within the sector through 'taxes', in the form either of a percentage of operators' revenue or of a tax on particular services, such as interconnection, traded at the wholesale level. The revenue thus raised is then distributed to ensure that the universal service obligation is satisfied. This could in principle be done either by subsidising particular groups of customers, to whom, for example, vouchers could be distributed, or it could be achieved by subsidising operators providing services to particular customers. The former model might seem particularly appropriate to cases where the obligation relates to particular categories of customers, but it is less suited to the case where a net cost of the USO arises because of geographical cost differences. As a result, the normal procedure is to impose the universal service obligation as a designated operator. 3. Regulating Broadband Arguably the most significant task regulators currently face concerns broadband, which is undergoing a crucial stage of the diffusion process where take-up accelerates and the service takes on a mass market dimension. As experience with mobile telephony shows, this can happen very quickly, and the competitive structure which emerges can be hard to dislodge as it is generally harder to take business from a competitor than to win a first-time customer. Broadband can be supplied using wire-based and wireless (fixed and mobile) technologies. In countries with cable networks capable of being digitised, a ready-made platform exists which can compete with the PSTN in the supply of broadband. Indeed, in some countries, cable is still the predominant form of delivering the service. Wireless technologies have so far made little impact. Wifi hotspots provide service in key areas but often as a radio ‘tail’ for a fixed service. Longer range fixed or nomadic wireless broadband technologies such as WiMax are still emerging. 3G mobile telephony has had difficulty in gaining revenue from data, rather than voice traffic. However, in areas where wireline networks are absent, wireless broadband has huge potential. -7TD 31 (PLEN/3) As a result, regulators often are confronted with a situation in which the key issue is whether the incumbent fixed operator’s monopoly or near-monopoly position in narrowband is going to be repeated in broadband, or whether the development of broadband creates an opportunity for competitors to build their own fixed infrastructures to compete with that of the incumbent. One framework in which to address these decisions is provided by the so-called ‘ladder of investment’ or ‘stepping stones’ theory, alluded to above, according to which competitors build up the capability to provide infrastructure-based competition by progressively replicating the incumbent’s assets, beginning with those that are relatively easy to replicate.12 As noted above, in the case of traditional narrowband voice telecommunications services, it is effectively impossible for an entrant to replicate call termination. The entrant has to gain access to the called party in order to complete the call, and this may only be done via the incumbent’s facility. In the case of call origination, replication will depend upon the nature of the entrant, for example a cable operator with a pre-existing local network. This example also illustrates the way in which ease of replication depends upon technological development: the availability of a wireless local loop would clearly ease the replication of call origination assets. Generally, the analysis leads to the not unexpected conclusion that the way to promote infrastructure competition is to make available easy and inexpensive access to the assets of the incumbent which are not replicable. At the outset this might include a large number of assets, which initially are complements to the entrant’s investment, but with time become substitutes. The entrant passes progressively through several stages of infrastructure competition as it ascends a ‘ladder’ of infrastructure construction. Next Generation Networks (NGNs) As broadband rolls out, operators will increasingly seek to develop their core networks in the direction of IP transport for all services, that is interchangeably for voice, data, video etc. In Europe, several operators are seeking to combine numerous pre-existing voice and data networks in a single IP core network. Many of them are also redesigning their access networks, in the direction of taking fibre closer to the customers’ premises than the local exchange or concentrator – specifically to the street cabinet (or node) which will then supply services to the small number of local customers using a copper ‘sub-loop’. . At upper levels of the network hierarchy, this convergence of traffic should make the market more contestable, as other operators can fill up their competing networks, particularly if they are successful in gaining broadband customers. However, other regulatory problems intrude. Most prominently, a new network may have different points of interconnection. In particular the equivalent of ‘local exchange’ interconnection may disappear, and as a result competitors’ infrastructure assets may be stranded. In the future they may have to match the incumbent by building out to the cabinet or node, or interconnect higher up the networks than the local exchange. This will alter the nature of competition based on access to the incumbent’s facilities. Regulatory policy to promote investment in NGNs is likely to depend on the competitive potential. In areas where there are both cable and telecoms networks, a race to build may develop. Where there is a fixed monopoly network, the government may need to offer subsidies or the regulator concessions in access pricing to get the new network built (Lewin, Williamson and Cave 2008). 12 See Cave (2006). -8TD 31 (PLEN/3) 4 Wireless Wireless networks, because of their cost structure, are subject much more easily to competition. Accordingly, they need little regulation apart from of termination rates, although their tendency to collude over pricing needs careful surveillance. The main intervention required is an active spectrum management which makes the maximum number of licences available for operators using the full range of available technologies (Cave, Doyle and Webb 2007). It is debatable whether the population is better served by mandating coverage requirements to achieve universal service goals or by relying on competitive interactions among operators (for one view on this, see Cave, Corkery and Tice 2008) Bibliography Buigues, Pierre A and Rey, Patrick (eds) (2004) The Economics of Anti-Trust and Regulation in Telecommunications, Edward Elgar, Cheltenham, Cave, Martin (2006) ‘Encouraging infrastructure competition via the ladder of investment’ Telecommunications Policy, (30) pp 223-237. Cave, Martin, Corkery, Matthew and Tice, Julian (2008) Competition and the Mobile Sector in Developed and Developing Countries, GSMA. Cave, Martin, Doyle Chris and Webb, William (2007) Essentials of Modern Spectrum Management, Cambridge University Press. Cave, Martin and Hatta, K. (2008) ‘Universal service obligations and spectrum policy’ INFO, 5/6, pp59-69. Fuss, Melvyn A. and Waverman, Leonard (2002) ‘Econometric Cost Functions’ in Cave, M. et al. (eds) Handbook of Telecommunications Economics Vol. 1., Elsevier, Amsterdam. ITU (2004) Trends in Telecommunications Reform 2004/5: Licensing in an Era of Convergence, Geneva. ITU (2008) Trends on Telecommunications Reform 2008: Six Degrees of Sharing, Geneva Lewin, David. Williamson, Brian and Cave, Martin (2008) Regulating Next Generation Fixed Access to Telecommunications Service, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1313704 OECD (2004) Access Pricing in Telecommunications, Paris. Sharkey, William W. (2002) ‘Representation of technology and production’ in Cave, M. et al, (eds) Handbook of Telecommunications Economics Vol. 1., Elsevier, Amsterdam. ______________